Successional mosquito dynamics in surrogate treehole and ground-container habitats in the northeastern United States: Where does Aedes albopictus fit in?

Authors


ABSTRACT:

This study assessed the risk of larval displacement of the eastern treehole mosquito, Aedes triseriatus, and the northern house mosquito, Culex pipiens, by Aedes albopictus, the Asian tiger mosquito, during the establishment and successional stages of novel larval mosquito treehole and ground-container habitats in the state of New Jersey, U.S.A. Culex pipiens and Culex restuans were the first mosquito species to colonize ground-container habitats and were the dominant larval species throughout the study period, whereas Ae. albopictus was late to colonize ground habitats and accounted for less than 15% of weekly larval collections once established. Ae. albopictus had a much stronger community presence within treehole ovitraps; however, Ae. albopictus never reached the average larval densities of the expected primary colonizer, Ae. triseriatus. Throughout the study period, the weekly abundances of Ae. triseriatus and Ae. albopictus were positively correlated and there were no significant differences between the abundances of each species. The larval dominance of Ae. triseriatus appears to be enhanced by the presence of Toxorhynchites rutilus septentrionalis, a large predatory mosquito species. When Tx. rut. septentrionalis was present, mature larvae (3rd–4th instar) of Ae. albopictus were also present in only 16.7% of collections, whereas mature larvae of Ae. triseriatus were collected concurrently with Tx. rut. septentrionalis in 53.8% of collections. These data suggest that Ae. triseriatus is at a greater risk of displacement by Ae. albopictus than are Cx. pipiens and Cx. restuans.

INTRODUCTION

The Asian tiger mosquito, Aedes albopictus, has been one of the most intensely studied mosquito species in the United States for nearly three decades since its introduction in Houston, TX, in 1985 (Moore et al. 1988). This invasive species has spread throughout most of the southeastern and eastern U.S. (Hawley et al. 1987, Moore 1999, Moore and Mitchell 1997) and has established itself as a major pest and potential vector species of several endemic viral diseases, such as West Nile virus (WNV), La Crosse encephalitis virus (LACV), and eastern equine encephalitis virus (EEE) (Mitchell et al. 1992, Mitchell et al. 1998, Turell et al. 2001, Holick et al. 2002). Globally, Ae. albopictus is a major vector of Dengue and Chikungunya viruses and is a potential vector of at least 20 other arboviruses (Moore and Mitchell 1997). In addition to being a prominent vector species, Ae. albopictus is currently the most invasive mosquito species in the world (Benedict et al. 2007). Ecologically, Ae. albopictus is a container-breeding mosquito species which oviposits in a variety of water-filled containers (both natural and artificial) and has been found to co-occur with larvae of several prominent container breeding mosquito species in the U.S., including Aedes aegypti, Culex pipiens, Aedes triseriatus, and Aedes japonicus (Bevins 2007, Carrieri et al. 2003, Juliano 1998). Numerous field and laboratory studies have documented the competitive superiority of larval Ae. albopictus over these and other mosquito species (Ho et al. 1989, Novak et al. 1993, Juliano 1998, Teng and Apperson 2000, Lounibos et al. 2001, Juliano et al. 2004, Braks et al. 2004, Bevins 2007). Two species in particular have been shown to be heavily impacted by the presence of Ae. albopictus in the U.S.; Culex pipiens and Aedes triseriatus (Teng and Apperson 2000, Costanzo et al. 2005).

Cx. pipiens and Ae. triseriatus are both abundant container-breeding species in the northeastern U.S. whose geographic distributions overlap with the northern limit of the current distribution of Ae. albopictus in the U.S. (Darsie and Ward 2005). In the Northeast, Cx. pipiens is the most abundant enzootic vector species for WNV and breeds prominently in small to large natural and artificial containers located on or near the ground (Carrieri et al. 2003, Hamer et al. 2008). Ae. triseriatus, better known as the eastern treehole mosquito, is the dominant vector of LACV in the U.S. (Haddow et al. 2009). Unlike Cx. pipiens, Ae. triseriatus breeds primarily in natural treehole habitats that are typically much smaller than the preferred habitat of Cx. pipiens (Leonard and Juliano 1995). The oviposition preferences of Ae. triseriatus and Cx. pipiens essentially eliminate any cohabitation and subsequent competitive interactions between the two species. Unlike either Cx. pipiens or Ae. triseriatus, Ae. albopictus breeds readily in a variety of natural and artificial containers of various sizes and is attracted to a wide range of oviposition attractants. This generalist response results in Ae. albopictus commonly cohabitating with both Cx pipiens and Ae. triseriatus. Larval competition studies involving Cx. pipiens, Ae. triseriatus, and Ae. albopictus demonstrate that the presence of Ae. albopictus can significantly decrease the survivorship, developmental time, and fecundity of both Cx. pipiens and Ae. triseriatus (Novak et al. 1993, Carrieri et al. 2003, Costanzo et al. 2005). However, despite the competitive superiority of Ae. Albopictus, neither of these native species have been displaced or have had their geographic distributions reduced by Ae. albopictus, as has been reported for other species (Hobbs et al. 1991, O'Meara et al. 1995, Kaplan et al. 2010).

The relative persistence of Cx. pipiens and Ae. triseriatus in the presence of the competitively dominant Ae. albopictus suggests that Cx. pipiens and Ae. triseriatus can co-occur successfully with Ae. albopictus. However, the mechanisms behind this co-occurrence are still unresolved, particularly regarding interactions that might occur outside the laboratory. One idea is that Cx. pipiens and Ae. triseriatus co-occur successfully with Ae. albopictus because each species occupies distinct temporal niches in relation to the establishment and occupation of their respective container and treehole habitats. Temporal niches may be established through nutrient fluctuation, water composition, and other extrinsic factors such as predation and parasitism. For example, the successful coexistence of Ae. triseriatus and Ae. albopictus in the southern United States appears to be a mediated by the predatory midge Corethrella appendiculata (Kesavaraju et al. 2008). However, field studies involving C. appendiculata have taken place solely in the southeastern U.S. where C. appendiculata, Ae. albopictus, and Ae. triseriatus commonly co-occur (Griswold and Lounibos 2006, Kesavaraju et al. 2008). It is unlikely that the influence of C. appendiculata would extend to other areas of the U.S. where the ranges of Ae. albopictus and Ae. triseriatus overlap, particularly in the northeastern U.S. where C. appendiculata is absent from treehole communities (Paradise 1998, 2004). However, similar patterns to those seen in the southern U.S. may be occurring in the Northeast as it is home to other treehole predators, such as Toxorhynchites rutilus septentrionalis, a northern subspecies of Toxorhynchites rutilus (Carpenter and LaCasse 1955), a common predator in treeholes in the southern U.S. (Lounibos et al. 2001, Griswold and Lounibos 2006). In regards to Cx. pipiens, much less work has been done to analyze extrinsic factors that may influence the competitive outcomes between Cx pipiens and Ae. albopictus. However, variations in resource type and habitat preferences appear to play a large role in the success of either species in the presence of the other (Carrieri et al. 2003, Costanzo et al. 2011).

Most of the above studies were conducted in a laboratory setting or involved the manipulation and stocking of artificial container habitats in the field. It is not clear that similar dynamics would occur in nature. For example, phenological differences between species, which are nonexistent in laboratory studies, can drastically alter the timing, intensity, and nature of the interactions occurring between species. The phenology, or seasonal pattern of abundance and activity, of each species is the product of a variety of mechanisms, including, but not limited to, temporal resource partitioning, predator avoidance, and physiological constraints (Morin 2011). Phenological differences not only determine the strength and nature of the interactions occurring between species (Ingold 1989, Lawler and Morin 1993, Snyder and Hurd 1995), but also dictate the assemblage patterns of seasonal mosquito communities inhabiting treehole and other inquiline habitats (Lounibos 1981, Bradshaw and Holzapfel 1984, Rango 1999). In order to identify the natural ecological mechanisms influencing the risk of displacement of Ae. triseriatus and Cx. pipiens by Ae. albopictus, we conducted a field experiment to analyze the seasonal patterns of establishment and succession of each species in novel larval ground-container and treehole habitats in central New Jersey (U.S.A.). Central New Jersey was chosen as the location of the study as it is within the northern limit of the current distribution of Ae. albopictus and other mosquito species, such as the predatory mosquito Toxorhynchites rutilus septentrionalis, which may result in unforeseen interactions.

MATERIALS AND METHODS

Study sites

Our study took place at two urban wetland sites located in Middlesex County, NJ (U.S.A.), during the 2011 season. Site one was Dismal Swamp (40°33′17.54″N, 74°23′11.82″W), a large protected urban wetland (>150 ha), and site two was Polansky Park (40°34′54.10″N, 74°21′24.73″W), a comparatively smaller urban wetland (<20 ha). These sites were chosen based on their ability to produce reliable catches of Ae. triseriatus, Cx. pipiens, and Ae. albopictus using CO2-baited CDC light traps (John Hock Company, FL U.S.A.) and CDC gravid traps in past field surveys (Johnson et al. 2012). Three trap sites were created between the two locations. Two sites were created within Dismal Swamp and one site was created within Polansky Park. Both sites located in Dismal Swamp were placed within the wetland 150 m from the wetlands' edge and were separated by 675 m. Site three was located in the center of Polansky Park 115 m from the wetlands' edge. The study began in the first week of July and ended during the first week of September. This 8 week time frame was chosen since all the species are readily abundant in our site locations over this time frame based on past adult mosquito surveys.

Container and treehole ovitraps

Two ovitraps (ground/treehole) baited with unique nutrient sources were used to select for Cx. pipiens and Ae. triseriatus larvae. Eight replicates of each trap type were set at each site for a total of 24 ground ovitraps and 24 treehole ovitraps. Each replicate container was assigned a number between 1–8 and was placed at random in a 2 by 4 grid pattern with at least 5 m between ground traps and 10 m between treehole traps. Each week a random ground trap and treehole trap from each site was chosen for collection. Ovitraps for Cx. pipiens were created out of 12-qt black containers. These containers are the default size of containers supplied with commercial gravid traps (John Hock Company, FL U.S.A.) that have proven successful at attracting gravid Culex spp. mosquitoes. The containers were baited with a hay infusion of 6 liters of distilled water and 30 g of timothy hay wrapped in a nylon sac; this is a strong ovipositing attractant for Culex spp. (Jackson et al. 2005). The hay infusion was created individually for each ovitrap in the field on the first day of the study. Ground traps were staked down and covered with 23-gauge galvanized wire mesh that prevented large animals from disturbing the traps but allowed for oviposition by mosquitoes and other insects.

Treehole ovitraps were created using black plastic 1 liter cups. To provide a surface for ovipositing Ae. triseriatus and Ae. albopictus mosquitoes, a wooden paint stir stick was placed into each container. Treehole ovitraps were baited with a white oak (Quecus blancas) leaf infusion consisting of 500 ml of distilled water and 8 g of crushed white oak leaves. A white oak infusion was used because white oaks are highly abundant within our study sites and infusions consisting of white oak leaves have been shown to be strong ovipositing attractants for both Ae. albopictus and Ae. triseriatus (Trexler et al. 1998). Treehole traps were hung at a height of 2 m above the ground on available white and red oak trees (Quercus alba, Quercus rubra) facing west.

Collection and identification

Ground traps were sampled using 250 ml mosquito sampling cups. After stirring to disorient and disperse mosquito larvae, four 250 ml samples were taken from each trap for a total of 1 liter. Distilled water was added to replace water taken, and water was added to all other traps to maintain a water level at 6 liters. The traps were then randomly shuffled to avoid confounding spatial effects. Treehole traps were collected by pouring out all contents into white enamel sampling containers and extracting all visible larvae and macroinvertebrates. The trap water was then poured back into the ovitrap. After the water was added back into the container, all traps were reshuffled at random to eliminate any spatially confounding effects. All collected larvae and pupae were placed into 50 ml glass containers and preserved in a 70% ethanol solution for later identification. Late instar (3rd–4th) larvae were identified using Darsie and Ward (2005).

Given the difficulty in correctly identifying early instar (1st–2nd) mosquito larvae, the majority of early instars were identified to genus. To identify the early instar larvae inhabiting treehole and ground-container ovitraps to species, subsamples of early instar larvae were allowed to mature to 4th instar for species identification. Paired t-tests were used to evaluate significant differences in the weekly abundances of species inhabiting ground and tree ovitraps. Correlations between species abundances were analyzed using Pearson product correlation coefficients. For statistical analysis, the data were square root transformed to more closely match assumptions of normality. All statistical analyses were done using R software (R Development Core Team 2008) and its related packages.

RESULTS

Ground ovitraps

Ground ovitrap communities were colonized first and were dominated by Culex spp. (Cx. pipiens and Cx. restuans) throughout the entire collection period (Table 1). Overall, Cx. restuans, a closely related species to Cx. pipiens with a similar geographical range (Carpenter and LaCasse 1955), was the first species to colonize ground-container habitats and was the dominant larval species throughout the study period, outperforming both Cx. pipiens and Ae. albopictus (Figure 1). Late instar Cx. restuans larvae accounted for 24.9±9.0% of weekly larval collections, whereas late instar Cx. pipiens larvae accounted for 13±2.9% of weekly larval collections. The difference between the larval densities of Cx. restuans and Cx. pipiens remained relatively constant throughout the entire study period. Late instar Cx. pipiens larvae maintained an average larval density equal to 50±15% of the density of late instar larvae of Cx. restuans when collected concurrently. Despite this difference, the weekly site abundances of Cx. restuans and Cx. pipiens were not significantly different (t22=1.375, P=0.18). Furthermore, there was a strong positive correlation (r=0.83, P=0.01) between the average weekly abundances of each species showing a strong synchronicity between the weekly abundances of Cx. restuans and Cx. pipiens.

Table 1.  Weekly ground and treehole ovitrap collection summaries. Numbers are represented as the total number of individuals collected per trap type for each collection week. Ground ovitraps were baited with a timothy hay-infusion (30 g/liter) and treehole ovitraps were baited with a white oak (Quercus alba) leaf infusion (6 g/500 ml).
Week/Species12345678
  Ground Ovitrap Summary (#/liter)
Culex restuans (3rd–4th instar)79143544026000
Culex pipiens (3rd–4th instar)03837921542
Aedes albopictus (3rd–4th instar)0003331611
Culex spp. (1st–2nd instar)227118695342500
Culex spp. pupae0192013610
Aedes spp. (1st–2nd instar)0000001724
Aedes spp. pupae0000072110
Anopheles punctipennis (3rd–4th instar)00364000
  Treehole Ovitrap Summary (#/500 ml)
Aedes triseriatus (3rd–4th instar)000104138102
Aedes albopictus (3rd–4th instar)0000351237
Aedes spp. (1st–2nd Instar)018235026211
Aedes spp. pupae0039411228
Toxorhynchites rutilus septentrionalis (3rd–4th instar)00110130
Figure 1.

Summary of late instar (3rd-4th instar) larval collections from ground-container habitats baited with a timothy hay infusion (Avg ± SE). The figure illustrates the seasonal dominance of Cx. restuans and Cx. pipiens over Aedes albopictus throughout the majority of the study period (weeks 1–7). Ae. albopictus was late to colonize ground-container habitats (week 4) and never overtook the larval densities of Cx. restuans and Cx. pipiens once established.

Unlike Cx. restuans and Cx. pipiens, Ae. albopictus did not colonize ground ovitraps until the midpoint of study period (week 4). Furthermore, once established, Ae. albopictus never reached the larval densities of either Cx. restuans or Cx. pipiens. Subsequent to colonizing ground-container habitats, late instar Ae. albopictus larvae only accounted for 11.3±6.1% of weekly larval collections, with significant differences (t22=1.86, P=0.015) occurring between the weekly site abundances of Cx. restuans and Ae. albopictus. Additionally, the weekly site abundances of Ae. albopictus were significantly negatively correlated with the abundance of Cx. restuans (r=−0.45, P=0.03) but not with the abundance of Cx. pipiens (r=0.14, P=0.52).

In regard to early instar larval densities, ground-container habitats were once again dominated by Culex spp. (Figure 2). Early instar Culex spp. larvae accounted for the overall majority of weekly larval collections (34.6±11.5%), which takes into account the total number of late and early instar larvae collected from ground habitats. Early instar Ae. albopictus larvae were not detected until late in the study (week 7), even after the detection of late instar Ae. albopictus larvae. Overall, early instar Aedes spp. larvae accounted for only 13.8±9.8% of weekly larval collections after establishment, however, despite the low presence of Ae. albopictus in ground-container habitats, there was a shift in early instar larval dominance late (week 7) in the season (Figure 2). The shift in dominance correlates with the first detection of early instar Ae. albopictus larvae. Anopheles punctipennis was also present in ground ovitraps between weeks 3–5, however, it never accounted for more than 2.0±1.3% of larval collections once present.

Figure 2.

Summary of early (1st-2nd instar) larval collections from ground-container habitats baited with a timothy hay infusion (Avg ± SE). The figure illustrates the seasonal dominance of early instar Culex spp. larvae belonging to Cx. pipiens and Cx. restuans over early instar Aedes spp. larvae belonging to Ae. albopictus over the first six weeks of the experiment. The figure also illustrates a late season succession from Culex spp. to Aedes spp. larvae within ground-container habitats demonstrating that Culex spp. may be at a greater risk of displacement from larval habitats by Ae. albopictus late in the season.

Treehole ovitraps

In contrast to ground ovitraps, Ae. albopictus had a much stronger community presence within treehole ovitraps but still never reached the larval densities of the expected primary colonizer, Ae. triseriatus (Figure 3). Once established, late instar Ae. triseriatus larvae accounted for an average of 25.2±8.6% of weekly collections, whereas late instar larvae of Ae. albopictus accounted for 13.6±5.4% of weekly collections after establishment. Despite these differences, the weekly site abundances of Ae. albopictus and Ae. triseriatus were not significantly different (t22=−1.29, P=0.20). Furthermore, the weekly abundances of Ae. albopictus and Ae. triseriatus were significantly positively correlated (r=0.66, P<0.001), demonstrating a strong synchronicity in the abundances of Ae. albopictus and Ae. triseriatus. Interestingly, the dominance of Ae. triseriatus may have been aided by the presence of the predatory mosquito Tx. rut. septentrionalis, which was the first late instar larval species collected in treehole ovitraps. Late instar larvae of Ae. albopictus were collected concurrently with the larvae of Tx. rut. septentrionalis in only 16.7% of weekly site collections, whereas late instar larvae of Ae. triseriatus were collected concurrently with Tx. rut. septentrionalis in 66.6% of weekly site collections. Unlike the larval densities of late instar Ae. albopictus and Ae. triseriatus larvae, early instar Aedes spp. larvae were apparently unaffected by the presence of Tx. rut. septentrionalis (Figure 4). Early instar larvae of Aedes spp. followed a symmetrical pattern of abundance that appears to be uninfluenced by the presence of Tx. rut. septentrionalis. This suggests that Tx. rut. septentrionalis may be preferentially feeding on larger late instar larvae.

Figure 3.

Summary of late instar (3rd-4th instar) larval collections from treehole habitats baited with a white oak leaf infusion (Avg ± SE). The figure illustrates that Ae. triseriatus was the first non-predatory mosquito species to colonize treehole habitats and, once established, maintained a higher weekly larval density than Ae. albopictus.

Figure 4.

Summary of early (1st-2nd instar) Aedes spp. and late instar (3rd-4th instar) Tx. rut. septentrionalis larvae collected from treehole habitats baited with a white oak leaf infusion (Avg ± SE). The figure illustrates that the seasonal abundance of early instar Aedes spp. larvae follows a symmetrical pattern that appears to be unaffected by the presence of the predatory mosquito Tx. rut. septentrionalis.

DISCUSSION

This is the first study to investigate the seasonal establishment and succession of novel treehole and ground-container habitats in the northeastern United States since the introduction of Ae. albopictus, the Asian tiger mosquito. Our data demonstrate the occurrence of a seasonal separation between the establishment and dominance of ground-container habitats by Culex spp. (Cx. restuans, Cx. pipiens) and Ae. albopictus. Culex restuans and Cx. pipiens established container habitats within the first week of the study, whereas it took an additional three weeks for Ae. albopictus to colonize ground ovitraps. This allowed for the seasonal dominance of Culex spp. within ground-container habitats. Although Ae. albopictus will oviposit within large to medium-sized ground-container habitats, from our data these appear not to be the preferred size for Ae. albopictus. Furthermore, there was a shift in early instar larval dominance from Culex spp. to Aedes spp. late in the season suggesting that it takes longer than expected for Ae. albopictus to become established within medium-sized larval habitats. Differences in oviposition activity between Ae. albopictus, Cx. restuans, and Cx. pipiens may be mediated through a variety of extrinsic factors, such as bacterial density, light intensity, and attractant type. For example, site specific attractants could have influenced the oviposition activity of Ae. albopictus in our study sites. Allan and Kline (1995) demonstrated that Ae. albopictus is more strongly attracted to larval rearing and field water than to hay infusions. Furthermore, Ae. albopictus has also been shown to oviposit in water sources containing certain bacteria species (i.e., Psychrobacter immobilis, Shingobacterium multivorum, Bacillus spp.) that were extracted from larval rearing water, soil-contaminated cotton towels, and an organic oak infusion (Trexler et al. 2003). Nonetheless, whatever the mechanism, our results demonstrate that Cx. pipiens and Cx. restuans can withstand displacement by Ae. albopictus within their preferred larval habitats, and this is at least partially mediated through temporal separation created by seasonal variation among species.

Compared to Cx. restuans and Cx. pipiens, Ae. triseriatus is at a much greater risk of displacement by Ae. albopictus in its preferred treehole habitats. There was a strong temporal synchronicity between the abundances of Ae. triseriatus and Ae. albopictus. This may be the result of both species preferring to oviposit in small container habitats typically containing leaf infusions (Leonard and Juliano 1995, Trexler et al. 1998). Furthermore, based on the competitive dominance of Ae. albopictus over Ae. triseriatus demonstrated in laboratory studies (Teng and Apperson 2000, Novak et al. 2003), there is strong evidence to suggest that Ae. triseriatus is at risk of displacement by Ae. albopictus in the northeastern United States. This in fact, may already be occurring. Rochlin et al. (2012) have shown that the statewide abundances of Ae. triseriatus in the state of New Jersey have decreased by a factor of three, while the statewide abundances of two major invasive mosquitoes, Ae. albopictus and Ae. japonicus, have increased by a factor of two over a nine-year period (2002–2011). This information, combined with demonstrated laboratory effects, provides strong evidence that Ae. triseriatus is at a high risk of displacement by invasive container-breeding mosquitoes in the Northeast.

Despite evidence that suggests Ae. triseriatus is at a high risk of displacement by Ae. albopictus, there is contradictory evidence to suggest that Ae. triseriatus may actually maintain its presence in the Northeast through the aide of native predatory species. In our study, the dominance of Ae. triseriatus in treehole habitats may have been influenced by the presence of the predatory mosquito Tx. rut. septentrionalis, which appeared to have a stronger negative impact on the presence and abundance of Ae. albopictus larvae than on larvae of Ae. triseriatus. If correct, these results would confirm the strong top-down effects induced by native larval predators shown to occur in treehole ecosystems in the southern U.S. (Bradshaw and Holzapfel 1983, Lounibos 1985, Kesavaraju et al. 2008). However, given the low number of Tx. rut. septentrionalis collected, seasonal variations in species abundance may be the biggest factor determining the differences in larval abundance between Ae. triseriatus and Ae. albopictus. In the Northeast, Aedes triseriatus typically reaches a peak in larval abundance in mid-to-late July, whereas Ae. albopictus typically reaches a peak in larval abundance in late July-August in temperate regions (Alto and Juliano 2001). Whatever the mechanism may be, the competitive interactions occurring within treehole ecosystems in the Northeast warrant further examination.

In conclusion, our results show that ground and treehole larval container habitats are dominated by their respective native colonizers (Cx. restuans, Cx. pipiens, Ae. triseriatus), and not by Ae. albopictus as expected from lab-based inferences (Teng and Apperson 2000, Costanzo et al. 2005). Studies investigating the impacts of non-native invasive species on native species have grown exponentially over the last two decades (Lockwood et al. 2005). The spread of Ae. albopictus, the most contemporary example of an elite invasive species, is a major public health concern because of its ability to transmit over 20 arboviruses. However, the majority of studies investigating the potential impact of Ae. albopictus on native mosquito species have been laboratory based. This study suggests that natural field studies can shed additional light on the competitive interactions occurring between Ae. albopictus and native container-breeding mosquito species occupying the same geographic range.

Acknowledgments

This work was funded by a Rutgers Department of Ecology and Evolution Small Grant awarded to BJ Johnson. We thank Mohammad Mallick and Viraj Dalal who participated in the field collection portion of the study. We also thank the two anonymous reviewers who provided constructive comments that helped us to improve the manuscript.

Ancillary